Professional Documents
Culture Documents
Editorial board
HIDE I S H I G U R O , KEIO U N I V E R S I T Y , J A P A N
ALAN MONTEFIORE, BALLIOL COLLEGE, OXFORD
MARY TILES, UNIVERSITY OF HAWAII
DAVID HOLDCROFT
Professor of Philosophy, University of Leeds
Published in the United States of America by Cambridge University Press, New York
www.cambridge.org
Information on this title: www.cambridge.org/9780521326186
A catalogue record for this publication is available from the British Library
EILEEN MARY
CONTENTS
Preface page ix
Introduction 1
1 Saussure's work: its context and significance 4
2 The distinction between langue and parole 19
3 Language as a system of signs, I: Signs,
arbitrariness, linearity, and change 47
4 Language as a system of signs, II: Diachronic and
synchronic linguistics 69
5 Language as a system of signs, III: Identities,
system, and relations 88
6 Language as a system of signs, IV: Values,
differences, and reality 107
7 Successes and failures 134
Notes 161
B ibliography 175
Index 178
Vll
PREFACE
Leeds
INTRODUCTION
General psychology
Individual psychology
[Etc.
Figure 1.1
first the object, then the sign; therefore (what we always deny),
that there is an external basis to the sign, and that language can
be represented as follows:
-a
objects * b names
* r
Langage
Figure 2.1
Langue Parole
Social Individual
Essential Contingent
No active individual role Active role
Not designed Designed
LANGUE AND PAROLE 21
Conventional Not conventional
Furnishes a homogeneous Furnishes a heterogeneous
subject matter for a branch subject matter studied by
of social psychology different disciplines
More will be said in the next section about each of these contrasts
and Saussure's reasons for making it, together with the distinc-
tion it underpins.
Figure 2.2
c = concept
s = sound image
Phonation Audition
Figure 2.3
2.1.3. Speech
It may seem odd that the discussion of the distinction between
langue and parole has taken so long to introduce the latter term,
but in this respect it is not unfaithful to Saussure's own discus-
sion. It is only in the course of the explanation of how it is that
the average conception of langue is crystallised in the brains of
members of the community that the term is introduced, almost
as an aside. After arguing that the psychological part of the
circuit is not wholly responsible for the crystallisation of lan-
guage, and that the executive component must be excluded, he
adds: 'Execution is always individual, and the individual is always
its master: I shall call the executive side speaking (parole)' (CLG,
30, 13). This suggests, given the explanation of'executive', that
the domain of speech is to be restricted to what goes on when
an acoustic image is associated with a concept, but that does not
seem to be Saussure's real view. For a little later he attributes a
much wider domain to it: 'Within the act, we should distinguish
between: (1) the combinations by which the speaker uses the
language code for expressing his own thought: and (2) the psy-
chophysical mechanism that allows him to exteriorize those com-
binations' (CLG, 31, 14). The students' notes make it clear that
by the former Saussure had in mind an individual's use of the
linguistic code to express thought, and by the latter everything
involved in speech production, including phonation. In other
words, the linguistics of parole has as its subject matter everything
in the speech circuit that is not passive; it is hardly surprising
that, if this is how it is conceived, it is heterogeneous in the
extreme.
Indeed, the subject matter of parole is even more extensive
than it seems at first sight. The brief passage on the subject gives
LANGUE AND PAROLE 31
no idea of the extent envisaged by Saussure of the individual's
freedom when making use of a linguistic code to express
thought. It is not simply a matter of a choice between alternative
formulations of what is intuitively the same thought, such as that
between:
and
and
Saussure Lyons
Acoustic image Expression element
'bet'
Figure 2.4
To which one might add: so would the work which made Saus-
sure's own reputation.
This interpretation is certainly confirmed by the reasons Saus-
sure gives for including a section on writing in a disquisition on
general linguistics. This is necessary because the linguist ought
to acquaint himself with the widest possible variety of languages.
However, 'we generally learn about languages only through writ-
ing' (CLG, 44, 23), and that is why it is necessary to consider
both its utility and its defects.
Moreover, this interpretation explains why Saussure was in-
terested in an ideal alphabetic form of writing which would make
possible a completely faithful representation of the acoustic im-
age (Harris 1987, 44).15 It also explains why Saussure thought
that the fact that we can assimilate the system underlying dead
languages does not constitute a serious counterexample to the
primacy he allots to speech sounds. For if a writing system is
parasitic on a phonic system and is a more or less accurate rep-
42 SAUSSURE
resentation of it, then we cannot reconstruct the underlying
langue without reconstructing the system of speech sounds it
utilises. So, though as a reconstruction of what motivated Saus-
sure this explanation has a more limited scope than Derrida's,
it seems to me to be the more plausible explanation, given that
Saussure did not wish to write off historical linguistics but rather
to argue for a sharp separation of synchronic and diachronic
issues.
That said, the existence of ideographic systems threatens to
undermine Saussure's position entirely. In these cases, Saussure
claims, the sign represents the whole word without representing
its internal structure, the classic example of such a system being
Chinese. However, since the ideogram may have a structure of
its own which depends on quite different principles, there seems
to be no warrant for saying that the ideogram represents the
spoken word rather than that it is a different way of representing
the same idea. The mere fact that there is a one-to-one corre-
lation between spoken and written forms does not entitle us to
say that the latter represent the former; there is such a corre-
lation in Lyons's model, for instance. Oddly enough, Saussure
admits as much: 'To a Chinese, an ideogram and a spoken word
are both symbols of an idea; to him writing is a second language,
and if two words that have the same sound are used in conver-
sation, he may resort to writing in order to express his thought'
(CLG, 48, 26). But this concession seems to be fatal to his overall
position.
(4) I will take the old men and boys to the bank.
Figure 3.1
But the only reason they give for coining the new terms is as a
corrective to the tendency to think of the acoustic image itself,
e.g., arbor, as the sign, rather than the union of it with a concept.
Inevitably, the reader is left with the impression that 'signified'
is a near-synonym for 'concept' and merely has different asso-
ciations, whereas, as we shall see, Saussure tries to elucidate the
notion of a signified in terms of the notion of a value, albeit a
very special kind of value arising from social usage (6.1). So the
suggested identification of a signified with a concept, and not
even a concept of a special kind, is, to say the least, unfortunate,
since the dangers of lapsing into the sort of nomenclaturist the-
ory that Saussure so objected to are clear.4
Arguably, the editors had no option but to retain the term
'acoustic image', since one of Saussure's central theses is that a
signifier is an acoustic image, though the 'is' in question is pre-
sumably not one of identity, since there are acoustic images which
are not signifiers. However, it might be argued that analogously
they were right to retain the term 'concept', since a signified is
a concept, albeit one of a special kind.
Certainly for the speech-circuit model to play any part in Saus-
sure's argument, a signified has to be able to fulfil the role al-
located to a concept in that model. So perhaps what we are
witnessing at this point is an attempt to reconcile two apparently
contrary tendencies in Saussure's thought, the psychologism im-
plicit in the speech-circuit model on the one hand, and the em-
phasis on the social character of language which leads him to
treat a signified as a special kind of value on the other. Perhaps
that tension is resolvable, but until it is resolved the mixture of
terminologies is undoubtedly confusing. To avoid such confu-
sion we should, at this stage, think of the new terms 'signifier'/
'signified' as theoretical terms whose content has to be explained
by reference to the role they play in a theory that has yet to be
developed, rather than as near-synonyms of the old pair 'acoustic
image'/'concept'.
52 SAUSSURE
Now, according to Saussure, linguistic signs have two 'pri-
mordial characteristics'. They are arbitrary, and their signifiers
are linear. It is to these two key theses that we now turn.
The term should not be taken to imply that the choice is left
entirely to the speaker (we shall see below that the individual does
not have the power to change a sign in any way once it has become
established in the linguistic community); I mean that it is un-
motivated, i.e. arbitrary in that it actually has no natural connec-
tion with the signified. (CLG, 101, 68)
Harris's points are well taken. However, they are not conclusive.
The expression 'their elements' could be taken to refer to pho-
nemes amongst other things, but it is certainly not necessary to
LANGUAGE AS A SYSTEM OF SIGNS I 59
take it so, particularly since Saussure did not think that the
sounds which realise signifiers literally constitute them. Instead,
Saussure could have been referring to the elements of a complex
signifier, e.g., 'twenty-one', whose elements would be simple sig-
nifiers. And, as far as I can see, the point about syllabic stress
would constitute a putative counterexample only if the stressed
syllable were a signifier, so that the stress modified what was
signified in some way.
Whether this is right or not, the crucial question is what role
the principle of linearity plays in Saussure's system. What he
himself says is, as we saw, quite explicit, namely that the crucial
account developed later of the relations which signifiers can en-
ter into (5.3) depends on it, though unfortunately he does not
say why he thinks this is so. A reasonable conjecture at this point
would therefore be that the principle is meant to explain why
signifiers can have these, and only these, relations. Certainly
without such an explanation there is a serious lacuna in his over-
all position, since there is no theoretical justification for his claim
that signifiers can enter into only the kinds of relations that he
describes.
If I am right, his argument is that because the medium in
which signifiers are realised is auditory, they can enter into only
the kinds of relations he describes later. And whatever one thinks
of such an argument, it is quite clear that an argument is needed
at this point. This is especially clear if one assumes that Saussure
has adopted a semiological approach, so that his aim is to derive
a detailed description of the nature of a linguistic system from
fundamental principles about signs (7.4).
But as well as a justificatory role, the principle has also a meth-
odological role, as Harris points out (1987, 77). For when we are
confronted with a complex signifier such as 'one hundred and
seventy-one', all that the principle of the Arbitrariness of the
Sign can tell us is that it is arbitrary, and that so are its constituent
signifiers. But it alone cannot tell us how to segment it into
constituent signifiers; to do this another principle is needed. So,
Harris argues, if the premiss is accepted that spoken language
is articulated as a concatenated succession of linked elements,
then 'by combining both Saussurean principles linguistics is im-
mediately provided with a method of analysis, which will consist
basically of segmenting any given catenary sequence by deter-
mining which arbitrary sets of consecutive elements in the se-
60 SAUSSURE
quence could constitute single signifiants and which could not'
(ibid., 77). And indeed one account that Saussure gives of lin-
guistic analysis does fit this description.11
However, if it is clear both what the role of the principle is,
and that such a principle is needed for the development of his
theory, it is very far from clear that the principle does the job
it is intended to do. For, to anticipate Saussure's account of the
'mechanisms' of language (5.3), only one of the relations he posits
has a linear character:
(1978, 100)
Time
Figure 3.2
Community
of
speakers
Figure 3.3
3.5. Summary
Saussure was root and branch opposed to nomenclaturism, the
view that a language consists of names whose function is to label
independently identifiable objects or ideas. In opposition to this
view, he insists that a language is a system of signs, and that the
signifying features of a sign depend on its intralinguistic relations
to other signs belonging to the same system rather than on its
extralinguistic relations to pre-existing objects or ideas. Clearly,
two key ideas need development and expansion, that of a sign
and that of a system, and it is with the first of these that this
chapter was concerned.
A sign is, for Saussure, a 'two-sided entity' consisting of the
union of a signifier and a signified. A signifier is an acoustic
image, so that, for example, the signifier of the word 'cat' is not
the sound made when it is pronounced, but an image of the
sound; it is thus a psychological entity. A signified is initially
identified with the concept associated with the signifier, but this
identification, I argued, is potentially confusing; we should treat
it at best as provisional and await Saussure's considered expli-
cation of it as a value arising from a system (6.1).
One primordial principle governing signs maintains that they
are radically arbitrary. This is so because they are unmotivated.
There is nothing about the world which makes one signifier any
more appropriate for its signified than any other. Saussure dis-
cusses two apparent counterexamples, onomatopoeia and inter-
jections, and has no difficulty in showing that these are not
serious counterexamples to the principle. However, he does not
LANGUAGE AS A SYSTEM OF SIGNS I 67
raise, let alone discuss, other counterexamples which are theo-
retically much more important, so his arguments for the prin-
ciple of the Arbitrariness of the Sign are, at this stage anyway,
quite inadequate.
A second primordial principle concerns signifiers. It maintains
that since what a signifier represents is auditory, it can only
represent a span of time, which has a linear character — hence
the name of the principle, the Linearity of the Signifier. Though
much less discussed than the principle of the Arbitrariness of
the Sign, this principle is, in Saussure's view, equally important,
since 'the whole mechanism of language depends on it' (CLG,
103, 70). He thought that this is so, I argued, because without
appeal to the principle of Linearity, his account of the way in
which a language is constituted by a system of syntagmatic and
associative relations would be powerless to answer the question
of why these and only these types of relations constitute a lan-
guage (5.3). However, I argued that the Linearity principle is
far too strong from Saussure's point of view, since it implies that
the only features that signifiers can represent are ones of tem-
poral succession. It thus excludes the very possibility of associ-
ative relations, which are non-temporal by definition.
Discussion of the question of why languages are both mutable
and immutable introduces a new and important dimension into
the discussion: time. There are, Saussure argues, many reasons
why languages should be resistant to change. Their arbitrariness
means that a debate whether to change them could have no
rational basis; they are extremely complex, and only a few ex-
perts understand them; since everyone is a language user, there
is the difficulty of getting everyone to change old habits. On top
of all this, another factor must be taken into consideration: the
effect of time. Language is inherited, so that we say 'man' and
'dog' simply because that is what our ancestors did; hence,
changes have to contend with solidarity with our traditions and
practices as well as the weight of the collectivity.
In view of all this, it might seem impossible for languages to
change, yet they do. How is this possible? Saussure's answer to
this question is much less detailed than was his answer to the
question of why languages are stable. There is no review of the
kinds of factors that might be responsible for change; instead
we get an assertion that, whatever its nature, change can only
have one kind of effect, a shift in the relationship of the signifier
68 SAUSSURE
to the signified. Moreover, Saussure mentions only one factor
which makes languages vulnerable to change: their arbitrariness.
So his explanation of why languages are vulnerable to change
is, by comparison with his explanation of why they are resistant
to it, very thin indeed.
Finally, having introduced the topic of time and stressed its
importance, Saussure is faced with the question of what theo-
retical distinctions are called for by the study of linguistic change.
This leads him to introduce the second great dichotomy, that
between synchronic and diachronic linguistics; this is the topic
of the next chapter.
LANGUAGE AS A SYSTEM
OF SIGNS
/ / : Diachronic and synchronic linguistics
between (1) the axis of simultaneities (AB), which stands for the
relations of coexisting things and from which the intervention of
time is excluded; and (2) the axis of successions (CD), on which only
72 SAUSSURE
one thing can be considered at a time but upon which are located
all the things on the first axis together with their changes.
D
(CLG, 115,80)
The reason why a concern with values calls for such a distinction
is, according to Saussure's own notes, because values always im-
ply a system (Engler 2, 178). The implication is that values can
be studied only on the axis of simultaneities because there is
nothing systematic about change in their case.
Nevertheless, Saussure concedes that some values can be stud-
ied on both axes; for instance, since the value of a plot of land
is related to its productivity, we can trace its value through time
provided that 'we remember that its value depends at each mo-
ment upon a system of coexisting values' (CLG, 116, 80). Saus-
sure's thought here seems to be that it is not unreasonable to
postulate some relationship between changes in productivity and
changes in value, so that if these do not form a system they are
nevertheless systematic.
But even if this is correct, the case of linguistics is different.
What makes it different from economics is, Saussure argues, the
fact that whilst economic values are not purely arbitrary, lin-
guistic values are. Because this is so, language is a system of pure
values 'determined by nothing except the momentary arrange-
ment of its terms' (ibid.).1 Moreover, the sheer complexity of
language, one of the facts mentioned to explain its relative sta-
bility (3.4.1), compels one to study it separately on each of the
two coordinates distinguished above.
After considering various names for these two different stud-
ies, Saussure opts for 'synchronic linguistics' as the name of the
study which describes linguistic phenomena on the axis of si-
multaneities, and 'diachronic linguistics' for the other study,
LANGUAGE AS A SYSTEM OF SIGNS II 73
which is concerned with the evolution of languages. But these
explanations are, to say the least, underdetermined. Though it
is clear that synchronic linguistics is concerned with systems of
values in ways in which diachronic linguistics is not, and that the
latter is concerned with evolution, not simply with change, whilst
the former is not, the precise differences between the subject
matter and methodology of the two sciences are at this point
very unclear. Moreover, since the two axes cross, one could be
forgiven for assuming that what we are dealing with is one and
the same thing studied from two different points of view, which
is, as we saw, certainly not what Saussure believes (Harris 1987,
89).
Before moving on to examples that Saussure gives to illustrate
the differences between the two sciences, three aspects of his
argument so far are worthy of comment. First, the claim that
linguistics differs from other sciences concerned with values by
being concerned with pure values rests on a much stronger ver-
sion of the autonomy of language than the arguments about the
arbitrariness of the sign have produced to date (3.2). For whilst
we are left to glean what exactly is the difference between a pure
value and other values, the root idea seems to be that whereas
the latter are at least partially determined by factors external to
the system to which they belong, the former is not. So if Saus-
sure's account of the linguistic sign is correct, the values asso-
ciated with it would be pure ones precisely because the sign is
radically arbitrary, that is unmotivated (3.2). However, as we
saw, Saussure's argument to show that linguistic signs are radi-
cally arbitrary fails to show that this is indeed so.
Second, it is not immediately clear why, even if the values in
question are pure ones, Saussure should not concede that lin-
guistics can study them 'historically', at least to the extent that
economics can, by comparing the synchronically determined
value of something at one time with its synchronically deter-
mined value at another; that is, by employing the method of
analysing and explaining change outlined above.
The final point is connected with this: that in differentiating
the historical study of economic values from that of linguistic
values, Saussure comes very close to maintaining that there is
no such thing as the history of changes of the latter, and a fortiori
no such thing as the history of changes in a language. For if
linguistic values are determined only by the momentary arrange-
74 SAUSSURE
ment of terms, there cannot be any systematic factors which
explain why one arrangement succeeds another. If there were,
then linguistic values would be determined by more than the
momentary arrangements in question. But even if such a view
is ultimately defensible, it is surely surprising.
Period B
Period B
AT
LANGUAGE AS A SYSTEM OF SIGNS II 8l
j Synchrony
( Language <
(Human) Speech < (^ Diachrony
(^ Speaking
(CLG, 139, 98)
But the diagram is something of a puzzle. It suggests that there
could be both a synchronic and a diachronic study of langue,
whilst parole itself is left completely out in the cold as the subject
matter of neither branch of linguistics.
It is even harder to avoid the conclusion that as a summary
of the preceding discussion the diagram is quite misleading when
it is viewed in the light of the remark, on the same page, that
'everything diachronic in language is diachronic only by virtue
of speaking (parole)" (ibid.). This is so, Saussure argues, because
the germ of all change is to be found in parole'.
Even so, I argued that the chess model highlights a serious prob-
lem. In chess, the moment a player makes his move a new state
of the board comes into being, and the time taken actually to
make the move has no significance. Hence, the transition from
one state to another is in effect instantaneous, so that at any
given time the game is in some determinate state or other. Now,
if the analogy with a language were exact, then the language
could be analysed neatly into successive language states, and the
time taken to effect a transition could be ignored. But clearly
this is not so, since a transitional period may last for a consid-
erable time, during which there is no uniform practice in the
area subject to change.
Saussure's implicit solution to the problem posed by the dis-
analogy is clear. In the case of two languages, one of which is
evolving rapidly and the other not, 'study would have to be
diachronic in the former instance, synchronic in the latter' (ibid.).
LANGUAGE AS A SYSTEM OF SIGNS II 87
In other words, successive language states, which can be studied
synchronically, do not necessarily succeed each other immedi-
ately; there may intervene periods of change which can only be
the subject of a diachronic study. Reverting to the chess analogy,
this would be to treat the time taken to make a move as of
significance, so that whilst the move is being made the game is
not in any particular state at all (Harris 1987, 105). But the
implications of what is said at this point are, to say the least,
difficult to reconcile with Saussure's overall position. For surely
if there are times at which a synchronic description of a language
is not possible, then it is hard to see what a description of the
langue internalised by its speakers could be. But one of Saussure's
central claims is that without langue there can be no parole. So if
no synchronic description of a language is possible, it is hard to
see how, on his principles, there could be a diachronic one either.
Further, Saussure writes as though the conception of a stable
state is, at best, a methodological fiction: ' . . . since language
changes somewhat in spite of everything, studying a language-
state means in practice disregarding changes of little importance,
just as mathematicians disregard infinitesimal quantities in cer-
tain calculations, such as logarithms' (CLG, 142, 101). But fiction
or not, the assumption that there are such states, which, more-
over, form the subject matter of synchronic linguistics, seems to
be central to Saussure's thought. However, it is far from clear
that he succeeded in developing a satisfactory model which dif-
ferentiates the subject matter of synchronic linguistics from that
of diachronic linguistics. Certainly the original model involving
only the two temporal axes is inadequate, whilst the analogy with
chess, which forms the basis for Saussure's preferred model,
limps badly at a number of points.
We shall return to some of these issues in Chapter 7. But we
must now turn to Saussure's detailed account of the way in which
a language constitutes a very special kind of system, one of pure
values.
LANGUAGE AS A SYSTEM
OF SIGNS
/ / / : Identities, system, and relations
88
L A N G U A G E AS A SYSTEM OF S I G N S III 89
Take French sizlaprd. Can we cut the chain after / and make sizl
a unit? No, we need only consider the concepts to see that the
division is wrong. (CLG, 146, 104)
In fact, Saussure goes on to argue, there are only two possible
divisions, both of which are determined by the meaning ex-
pressed by the words: si-z-la-prd and si-z-l-aprd - respectively, si
je la prends ('if I take it') and sije Vapprends ('if I learn it').
One obvious consequence of this proposal is that concrete
L A N G U A G E AS A S Y S T E M OF S I G N S III 91
Saussure hastens to add that this does not make the train an
abstract entity, since it must have a material realisation. Though
there is no particular group of carriages and so forth that it must
consist of, nevertheless it must consist of some. Thus the identity
of the successive trains cannot be accounted for in terms of the
material entities of which they are composed, for these might
be completely different. What makes the two trains the same, in
spite of their material differences, is the fact that each is related
in the same way to a time of departure, a route, a set of con-
nections with other trains, etc.
Another analogy that he develops to illustrate his point in-
volves yet another comparison with chess:
(i) That there are two generic kinds of relation that signifiers
and signifieds can enter into, syntagmatic and associative
(ii) That the specific kinds of syntagmatic and associative rela-
tions which relate signifiers to signifiers are different from
those which relate signifieds to signifieds
(iii) That the identity of a particular signifier or signified is purely
differential
(iv) That a system of signs results from the pairing of a system
of signifiers with one of signifieds; moreover, signs as such
have a value and are not purely differential entities
I shall discuss the first two points in the remainder of this
section and the other two in the next chapter.
un-do
un-do —
un-glue do
\
un-pick re-do
un-dress etc.
/ \
etc.
lui
Que
me
dit-il?
vous
nous (Engler 2, 294E)
In reality the idea evokes not a form but a whole latent system
that makes possible the oppositions necessary for the formation
of the sign. By itself the sign would have no signification. If there
were no forms like marche! marchez! against marchons!, certain
oppositions would disappear, and the value of marchons! would
be changed ipso facto. (CLG, 179, 130)
To frame the question que vous dit il? 'What does he say to you?'
the speaker varies one element of a latent syntactical pattern, e.g.
que te dit il? 'What does he say to thee? que nous dit il? 'What does
he say to us7 etc., until his choice is fixed on the pronoun vous.
In this process, which consists of eliminating mentally everything
that does not help to bring out the desired differentiation at the
desired point, associative groupings and syntagmatic patterns
both play a role. (CLG, 180, 130)
LANGUAGE AS A SYSTEM
OF SIGNS
IV: Values, differences, and reality
Figure 6.1
'in language there are only differences without positive terms.
Whether we take the signified or the signifier, language has
neither ideas nor sounds that existed before the linguistic system,
but only conceptual and phonic differences that have issued
from the system' (CLG, 166, 120).
'Signified ^ ^Signified
whether there is any work left for the latter term in Saussure's
theory, since if its use arises from adopting a partial point of
view it is hard to see how it could be anything other than
misleading.
However, to understand Saussure's answer to this question,
we must look at the two not obviously compatible metaphors to
which he appeals to explain what a value is. According to the
first, values are like units of exchange; according to the second,
they are the products of the interaction of two otherwise undif-
ferentiated substances, thought and sound.1
6.3. Thought—sound
The second metaphor involves a contrast between form and
substance, in which a langue is likened to a form which differ-
entiates otherwise amorphous substances. Considered in abstrac-
tion from language, ideas are, Saussure argues, amorphous, so
that 'nothing is distinct before the appearance of language' (CLG,
155, 112). Further, before the appearance of language, sounds
do not 'yield predelimited entities' - a conclusion which is hardly
surprising in the light of the earlier discussion of the French
word nu (2.1.1). The linguistic fact therefore has to be pictured3
'as a series of contiguous subdivisions marked off on both the
indefinite plane ofjumbled ideas (A) and the equally vague plane
of sounds (B). [Figure 6.3] gives a rough idea of it.' (CLG, 156,
112). So the effect of language on thought is not to create a
LANGUAGE AS A SYSTEM OF SIGNS IV
Figure 6.3
Wfsheit
Kunst List
ca. 1300
(Lehrer 1974, 15)
They find that all languages have terms for black and white. If
there is a third term, it will be red. If a language has four terms,
ll8 SAUSSURE
the fourth will be either yellow or green. Languages with five
terms have both yellow and green. A word for blue is the sixth
to emerge, and a term for brown is the seventh. If a language
has eight or more colour words, it will have words for purple,
pink, orange, gray, or some combination. (Lehrer 1974, 153)
Thus we can say that though phonemes are values, since they
are determined by the reciprocal oppositions in which they stand
to other phonemes, they do not have values, since they signify
nothing.
But, as Jakobson urges, the characteristic of phonemes of
being purely negative, of lacking content, as well as of being
relative and opposing entities, is not a characteristic of other
linguistic entities, as Saussure claims. For instance, Saussure ar-
gues that letters are, like phonemes, purely negative and dif-
ferential:
The only requirement is that the sign for t not be confused in his
script with the signs used for Z, df etc.
LANGUAGE AS A SYSTEM OF SIGNS IV 123
(3) Values in writing function only through reciprocal opposition
within afixedsystem that consists of a set number of letters. This
third characteristic, though not identical to the second, is closely
related to it, for both depend on the first. (CLG, 165, 119)
although both the signified and the signifier are purely differ-
ential and negative when considered separately, their combination
is a positive fact; it is even the sole type of facts that language
has, for maintaining the parallelism between the two classes of
124 SAUSSURE
differences is the distinctive function of the linguistic institution.
(CLG, 166, 120)10
In other words, though it is not true that the only role that
signifiers have is a differential one (because of the correlations
that exist between them and signifieds), the fact remains that,
this correlation apart, they are constituted only by reciprocal
oppositions within a system. Hence, if we consider them only
from this point of view, they are purely negative and differential
entities, as indeed are signifieds. It is to the consideration of this
claim that we now turn.
6.4.3. Differences
Because of its intrinsic interest and philosophical implications,
we shall restrict our discussion to the case of signifieds. The claim
that signifieds are purely negative and differential entities has
arguably two different sources in Saussure's thought. One
source, connected with the metaphor of the amorphous pre-
linguistic masses of thought and sound, argues that the negative
and differential role of signifieds arises from the fact that they
occupy, or correspond to, an arbitrarily chosen part of a contin-
uum which has no positive features to distinguish it essentially
from other arbitrarily chosen parts. Culler, for instance, sees this
as an important source of Saussure's ideas:
Differences
Oppositions
Figure 6.4
6.5. Summary
In this chapter we have been tracing the ways in which the
threads of Saussure's complex argument come together. In par-
ticular we have been concerned first with the connection he
makes between the notion of value and that of system, and sec-
ond with his claim that values are relative, opposing, and negative
entities, so that in languages there are only differences without
positive terms.
Since values are products of a system, it is important to note
that the theoretical implications of the term are radically differ-
ent from that of the term 'signification' which it supplants, since
they are holistic (3.1.1); we cannot first identify a value and then
describe the system to which it belongs, since 'words present them-
selves as terms of a system' (SM, 90). To illustrate the way in which
LANGUAGE AS A SYSTEM OF SIGNS IV 131
values depend on a system, Saussure appeals to two metaphors.
In the economic metaphor he argues that, outside of language,
values have two features (6.2). There are dissimilar things for
which they can be exchanged, and similar things with which they
can be compared. For instance, a five-franc piece can be exchanged
for a loaf of bread and compared with a one-franc piece, a ten-
franc piece, etc. Analogously, a word can be exchanged for an
idea or compared with another word. And just as one can grasp
the value of a coin only if one knows what it can be exchanged
for and how it is related to other coins of different denomina-
tions, so one can grasp the value of a term only if one knows
both what it signifies and the relations that obtain between it and
terms to which it contrasts.
However, there are a number of respects in which the com-
parison breaks down. Apart from the obscurity of the notion
that words can be exchanged for ideas, the analogy says that
what would have to be exchanged are tokens of signifiers, so
chat the transactions would belong to parole and not langue. On
the other hand, comparisons would have to be between entities
corresponding to units of currency, which are signifiers, and not
their tokens; but it is difficult to see how such comparisons could
account for the values that signifiers have, that is the signifieds
with which they are associated.
In addition, the analogy has a very seriousflawfrom Saussure's
point of view. The goods for which coins are exchanged have
values other than monetary ones which, at least in part, deter-
mine their monetary values, so that it is difficult to see how the
latter could be a product of the system of currency alone. But
if they are not, then they cannot be pure values, and as such
radically arbitrary.
In this respect the second of Saussure's metaphors, that of the
interaction of two amorphous masses, wind and water, is much
better from his point of view (6.3). For the system of differences
arising from their interaction, the waves, has no existence in-
dependently of it. But there are other difficulties with this com-
parison. The diagram illustrating the analogy suggests that
signifiers should be construed as slices of sound - contrary, of
course, to Saussure's view that they are acoustic images (3.1.1).
And whilst the waves are a product of the action of the air on
the water, Saussure does not wish to allot an analogous role to
the thought mass. If language is a form (2.2), then we need some
132 SAUSSURE
account of the way in which 'something' which is neither thought
nor sound provides a principle for structuring both. And if at
this point to provide such an account we need to appeal to the
chess analogy, as I suggested, then the thought-sound metaphor
simply does not stand on its own feet.
As for the argument that, even so, the metaphor does draw
attention to an important truth - that languages provide prin-
ciples for making distinctions within phenomena which present
themselves as undifferentiated continua, e.g., the colour spec-
trum - I argued that it too is wanting (6.3.1). Apart from the
fact that the colour spectrum itself is hardly an object of everyday
experience, casting doubt on the literal applicability of argu-
ments that something which presents itself as undifferentiated
is then differentiated, the work of Berlin and Kay strongly sug-
gests that colour vocabularies are by no means completely ar-
bitrary. So, since Saussure's influential metaphors shed so little
light, I tried to describe the way in which his argument develops
independently of them, by tracing the relation between a value
and a system and the reasons for thinking that values are op-
posing, relative, and negative entities (6.4).
The connection between a value and the system to which it
belongs is, I argued, governed by the principle (T) that the value
of a linguistic item is determined by the set of syntagmatic and
associative relations that it enters into with other items in a langue.
Moreover, for Saussure not only are signifieds values, but so are
signifiers and phonemes (in the modern sense of that term),
though they are, of course, values of different kinds.
However, though signifiers, signifieds, and phonemes all are
values, the latter certainly do not have a value, since they are not
correlated with signifieds (6.4.1). Further, the distinction be-
tween being a value and having one is important for the eval-
uation of the claim that values are purely negative. One thing
that Saussure might have meant by this can be illustrated by
reference to phonemes. The important point is that for a given
phoneme to play its linguistic role — which, in Jakobson's words,
is 'to distinguish the word containing this phoneme from any words which,
similar in all other respects, contain some other phoneme' (1978, 62) —
it is necessary only that it be different from the other phonemes
in the language. It is not also necessary that it have a positive
content, i.e., that it should signify something. But if this is what
Saussure meant, it is true neither of letters nor of signifiers
LANGUAGE AS A SYSTEM OF SIGNS IV 133
generally. They do not have a purely differentiating role, for as
well as being values they have values. Significantly, we noticed
that when talking about letters Saussure maintained that the
character of being purely negative is closely associated with that
of being relative and opposing, even though letters have the
latter characteristic but not the former, so that the association
cannot be logical. Clearly, phonemes cannot be taken as a model
of signifiers.
Finally, we considered the objection that Saussure could hardly
have denied that signifiers have values; after all, a langue is a
system for the correlation of signifiers and signifieds. That
granted, the objection continues, he maintains that, this corre-
lation apart, they are constituted only by reciprocal oppositions
within a system, so that if we consider them from this point of
view alone, they are indeed purely differential and negative, as
are signifieds.
However, we did not succeed in making sense of this claim in
the case of signifieds (6.4.3). Differences as such do not give rise
to oppositions; it is only in those cases in which signifieds (or
what they denote) belong to what is intuitively the same domain
that they do. Furthermore, though oppositions involve differ-
ences, they standardly involve similarities; as Lyons says, 'op-
positions are drawn along some dimension of similarity' (1977,
286). But why should the specification of the similarity not count
as a positive feature?
The binary thesis suggested an explanation of what it is for a
term to be negatively characterised - it is the unmarked term in
a binary opposition. But clearly this explanation requires other
terms to be positively characterised, other difficulties with the
proposal apart. Further, I argued that there are a number of
other important semantic relationships which call for a posi-
tive characterisation, e.g., kind membership and part—whole
relationships.
Thus not only did we fail to make sense of the thesis that
signifieds are purely differential and negative entities, but there
is good reason to think that this cannot be done. However, it
does not follow that Saussure's overall project shipwrecks. For
whilst the theory and practice of structural linguistics appeals to
the principle (T), it does not appeal to the principle (Ti), that
in language there are only differences without positive terms.
Moreover, (T) does not imply (Ti).
SUCCESSES AND FAILURES
... one is forced to admit that these specific facts reside in the
society itself that produces them and not in its parts - namely its
members. In this sense therefore they lie outside the consciousness
of individuals as such, in the same way as the distinctive features
of life lie outside the chemical substances that make up a living
organism. (Durkheim 1982, 39)
161
162 NOTES TO P P . 1 5 - 1 6
image and a concept, even though the latter term has individualistic
psychological connotations which are very misleading from the point
of view of the theory that Saussure goes on to develop, and quite
apart from the fact that the term has no intrinsic connections with
the term 'system'. It is true that later the new terms 'signifier' and
'signified' are introduced by the editors to replace the pair originally
used.
Ambiguity would disappear if the three notions involved here
were designated by three names, each suggesting and opposing
the others. I propose to retain the word sign [signe], to designate
the whole and to replace concept and sound-image respectively
by signified [signifie] and signifier [signifiant]: the last two terms
have the advantage of indicating the opposition that separated
them from each other and from the whole of which they are
parts. (CLG, 99, 67)
This passage suggests that, apart from the advantage of an implied
relationship, the two new terms have close family relationships with
the pair they replace, so that a signified is a concept. But in fact the
new terminology has radically different theoretical implications from
the old (6.1).
6. Examination of the sources reveals that this is yet another place at
which the initiative of the editors has been considerable. Apart from
the fact that the italicisation is their responsibility, for some reason
they discard the metaphor of the study of langue as a platform from
which one can view the position of other aspects of language (Ian-
gage). The metaphor is found in the notes of all the students. For
instance, Madame Sechehaye writes: '[It is] when one accords first
place to la langue, making it the point of departure, that one can
give their true place to the other elements of language (langage)'
(Engler 1, 30B). This certainly makes the study of la langue central,
in that it is the foundation of any kind of study of language, but
there is no suggestion that it is its own raison d'etre 'in and for itself.
So the italicised passage read in relation to its sources can hardly be
used to underwrite the interpretive summary of the whole of CLG
encapsulated in the final remarks.
7. As de Mauro points out, this order of presentation, which stresses
the interdependence of the study of languages and of language itself,
corresponded to a deep conviction of Saussure's which he articulated
in his 1891 inaugural lecture at Geneva:
... the most elementary linguistic phenomena will not be sur-
mised or clearly perceived, classified, and understood unless
one resorts in the first and last instance to the study of lan-
guages. ... On the other hand, to want to study languages ig-
noring the fact that they are governed primordially by certain
principles which are summarised in the idea of language (Ian-
gage) is a task even more bereft of any serious significance, of
any genuine scientific foundation. (TM, 354)
8. Here is a detailed comparison of the relevant passages:
NOTES TO P. 16 163
175
176 BIBLIOGRAPHY
Amacker, R. 1975. Linguistique saussurienne. Geneva and Paris: Librairie
Droz.
Barthes, R. 1967. Elements of Semiology, translated by A. Lavers and C.
Smith. New York: Hill and Wang.
Benveniste, E. 1966. Problemes de linguistique generate. Paris: Gallimard*
Berlin, B., and P. Kay. 1970. Basic Color Terms. Berkeley and Los An-
geles: University of California Press.
Burge, T. 1975. 'On Knowledge and Convention'. Philosophical Review
84, 249-55.
Calvet, L. 1975. Pour et contre Saussure. Paris: Payot.
Caws, P. 1988. Structuralism: The Art of the Intelligible. Atlantic Highlands,
N.J.: Humanities Press.
Chomsky, N. 1957. Syntactic Structures. The Hague: Mouton.
1964. Current Issues in Linguistic Theory. The Hague: Mouton.
1965. Aspects of the Theory of Syntax. Cambridge, Mass.: MIT Press.
1966. Cartesian Linguistics: A Chapter in the History of Rationalist Thought.
New York: Harper and Row.
1969. Language and Responsibility. Hassocks: Harvester.
Culler, J. 1976. Saussure. Glasgow: Fontana/Collins.
D'Agostino, F. 1986. Chomsky's System of Ideas. Oxford: Clarendon Press.
Danto, A. C. 1973. 'Methodological Individualism and Methodological
Socialism'. In O'Neill 1973, 312—37.
Derrida, J. 1976. OfGrammatology, translated by G. C. Spivak. Baltimore:
Johns Hopkins University Press.
1981. Positions, translated and annotated by A. Bass. London:
Athlone.
Dewey, J. 1958. Experience and Nature. La Salle, 111.: Open Court.
Doroszewski, W. 1933. 'Quelques remarques sur les rapports de la so-
ciologie et de la linguistique', Journal de Psychologie.
Ducrot, O. 1968. Le structuralisme en linguistique. Paris: Seuil.
Ducrot, O., and T. Todorov. 1972. Dictionnaire encyclopedique des sciences
du langage. Paris: Seuil.
Durkheim, E. 1982. The Rules of Sociological Method, edited with an
introduction by S. Lukes; translated by W. D. Halls. London:
Macmillan.
1952. Suicide: A Study in Sociology, translated by J. A. Spaulding and
G. Simpson. London: Macmillan.
Firth, J. R. 1957. Papers in Linguistics. Oxford: Oxford University Press.
Gibson, R. F. 1982. The Philosophy ofW. V. Quine. Tampa: University
of South Florida.
Giddens, A. 1978. Durkheim. Hassocks: Harvester.
Gochet, P. 1986. Ascent to Truth: A Critical Examination of Quine's Philos-
ophy. Munich: Philosophia Verlag.
Goffman, E. 1972. Relations in Public. Harmondsworth: Penguin.
Harris, R. 1987. Reading Saussure. London: Duckworth.
BIBLIOGRAPHY 177
Henry, A. 1970. 'La linearite du signifiant'. In J. Dierickx and Y. Lebrun
(eds.), Linguistique contemporaine: hommage a Eric Buyssens. Brussels:
Editions de L'Institut de Sociologie, 86—92.
Jakobson, R. 1978. Six Lectures on Sound and Meaning, translated by J.
Mepham. Hassocks: Harvester.
Lehrer, A. 1974. Semantic Fields and Lexical Structure. Amsterdam: North-
Holland.
Leiber, J. 1978. Structuralism: Skepticism and Mind in the Psychological
Sciences. Boston: Twayne.
Lepschy, G. C. 1970. A Survey of Structural Linguistics. London: Faber
and Faber.
Levi-Strauss, C. 1977. Structural Anthropology, translated by C. Jacobson
and B. Grundfest Schoepf. Harmondsworth: Penguin.
Lewis, D. 1975. 'Language and Languages'. In K. Gunderson (ed.),
Language, Mind and Knowledge. Minnesota Studies in the Philosophy
of Science, VII. Minneapolis: University of Minnesota Press, 3-35.
Lukes, S. 1973. Durkheim. London: Allen Lane.
Lyons, J. 1968. An Introduction to Theoretical Linguistics. Cambridge: Cam-
bridge University Press.
1977. Semantics, vol. 1. Cambridge: Cambridge University Press.
MacDonald, G., and P. Pettit. 1981. Semantics and Social Science. London:
Routledge and Kegan Paul.
Mill, J. S. 1875. A System of Logic, 9th ed. London: Longman.
Mounin, G. 1968. Saussure ou le structuralismesans lesavoir. Paris: Editions
Seghers.
O'Neill, J. (ed.). 1973. Modes of Individualism and Collectivism. London:
Heinemann.
Robey, D. (ed.). 1973. Structuralism: An Introduction. Oxford: Clarendon
Press.
Quine, W. V. i960. Word and Object. New York: Wiley.
1969. Ontological Relativity and Other Essays. New York: Columbia Uni-
versity Press.
1977. 'Facts of the Matter'. In R. S. Shahan and C. Swoyer (eds.),
Essays in the Philosophy of Quine. Hassocks: Harvester, 155-69.
Sampson, G. 1980. Schools of Linguistics: Competition and Evolution. Lon-
don: Hutchinson.
Schleicher, F. 1869. Darwinism Tested by the Science of Language, translated
by A. Bikkers. London: Hotten.
Trier, J. 1931. Der deutsche Wortschatz im Sinnbezirk des Verstandes. Hei-
delberg: Winter.
1934. 'Das sprachliche Feld', Jahrbuch fur Deutsche Wissenschaft 10.
Watkins, J. 1973. 'Ideal Types and Historical Explanation'. In O'Neill
Index
178
INDEX '79
Harris, R., 41-2, 48, 58-9, 163, 166, Mill, J. S., 147-8
169 Mounin, G., 5, 8, 144
historical linguistics, see comparative
linguistics Neo-Grammarians, 9
Hjelmslev, L., 44, 137 nomenclaturism, 11-12, 48-50, 66,
113, 142
identity, synchronic, 22, 92, 95-7,
104-5, 1X9 onomatopoeia, 54
interjections, 54 oppositions, linguistic, 110, 112, 121,
125-7; binary, 127-9
Jakobson, R., 52, 58, 60, 121-3, 132,
164-5 parole, 30-4, 84-5, 87, 138; and
langue, 20-1, 45-6, 137-9
Kay, P., 117, 132 philosophers, Saussure's advice to,
language, 4, 22-3, 134-5; autonomy phonemes, 120-3
of, 10-13, 17, 55, 65; compari- Popper, K., 151
son with chess, 76—8, 114—15, Port Royal Grammar, 10—12, 17, 55,
139, 168-9; comparison with 142, 166
Morse Code, 34, 37, 44-5; com-
parison with a symphony, 33, 37, Quine, W. V., 140-3
44—5; faculty of, 23-4, 42-4,
168; and social facts, 8, 22, 27- Rae,J., 168
30, 61-3, 65, 145-55; a n d Riedlinger, A., 13
thought, 112-15; see also abstract
and concrete entities of language Schleicher, F., 7-9
langue, 15, 21-4, 26-30, 34-6, 44-5, Sechehaye, A., 1, 13-16, 51, 136-7,
65—6, 84—7, 151—2; and associa- 161—2, 170
tive and syntagmatic relations, semiology, 5-6, 17, 19, 35-6, 59, 61,
100—4; a s a form, 36-9, 112-15; 79> 83> 94> 98> i37» 155~Qo
and parole, 20—1, 33—4, 45—6, signified, 14, 50—5, 108—11, 119, 162;
and differences, 121-30
Lehrer, A., 116, 128 signifier, 14, 50-5, 56-61, 165; see
Lepschy, G., 169 also linearity of the signifier
Levi-Strauss, C , 5, 53, 102, 158 signs, linguistic, 49—52, 155—60, and
Lewis, D., 153-4, 173-4 see arbitrary nature of the sign;
lexical fields, 116-17, 172 form a system, 93-7, 108-10,
linearity of the signifier, 56—61, 89, 119, 125, 135; immutability of,
94, 98, 137, 159 62-3; mutability of, 63-6; see also
linguistics, 15, 17, 73, 134-7, 160; value: linguistic
history of, 6-9; object of, 19-20, social facts, 8, 27-8, 61-3, 135, 145-
22-3, 34-6, 46, 81, 84-5, 164; 7; and Saussurean Linguistics,
see also comparative, diachronic, 152-5
structural, and synchronic social realism, 143, 148; and Saus-
linguistics sure, 152-5
Lukes, S., 146 speech sounds, 21-2, 40-3, 89,
Lyons, J., 38, 94, 125-7, 130, 133, 164-5
164, 170, 172 structural linguistics, 104, 130,
170
Mauro, T. de, 14-15, 162, 164 structuralism, 5, 130, 135
Meillet, A., 144 synchronic linguistics, 17, 69-71, 74-
methodological individualism, 147—9; 8, 81-5, 134-5, 139; is really
and Saussurean linguistics, 149- idiosynchronic, 82; see also iden-
52, 154-5 tity, synchronic
180 INDEX
syntagmatic relations, 57, 60, 83, 9 8 - 95~7> 119» x 21-30; and ex-
100; relations to langue and pa- change, 108-12; and the indis-
role, 100-4 tinctness of pre-linguistic
thought, 112-15; kinds of,
Tarde, G., i44~7> ^ 3 120-1; see also oppositions,
thought-sound, 112-15,117,124 linguistic
translation, 110; radical, 140-1
Trier, J., 116
Watkins,J., 148
value: linguistic, 51, 71-4, 77-9, 104, Whitney, W. D., 8, 23, 144
108—9, 12i—4; and differences, writing, 39-42, 156